feat: add Project Health Check (six dbt-project-evaluator dimensions)

[il-dat]

What & why

Adds a Project Health Check to dbdocs, derived from the static dbt artifacts
we already load — no extra dbt package, no warehouse. The manifest drives a rule
engine covering the six dbt-project-evaluator
dimensions (testing, modeling, documentation, structure, performance,
governance); an optional run_results.json adds per-test pass/fail detail. The
SPA surfaces a scorecard + collapsible dimensions page and per-model test
results on each node page. It's always built and fail-soft: a
missing/malformed run_results.json never sinks generate, and the SPA only
surfaces the section when it holds findings.

This PR also carries the supporting tooling: .github PR + issue templates, a
/dbdocs-pr skill/command, and a new UX & accessibility dimension on
dbdocs-code-review.

Type of change

• 🐞 Bug fix (non-breaking change that fixes an issue)
• ✨ Feature (non-breaking change that adds capability)
• 💥 Breaking change (existing behavior changes)
• 🧹 Refactor / internal (no user-facing change)
• 📖 Docs only

Area

• CLI (generate / serve / deploy) — new --run-results override
• extract/ (nodes / erd / graph / column_lineage / health) — new extract/health/ sub-package
• site/ (data dict / builder / 3-tier SPA / deploy) — health key + SPA pages
• frontend/ (React Flow graph bundle) — DAG type/schema filter round-trips through the hash
• Config (dbdocs.yml) / packaging / CI / docs — run_results/health knobs, artifact-parser dep, docs

How to test

# A plain `dbt build`/`dbt test` produces target/*.json incl. run_results.json.
task generate            # health section is built from the committed jaffle_shop fixtures
task serve               # then open http://localhost:8080/
#   → #/health           the scorecard + six collapsible dimensions
#   → any model page      the "Tests" section (data + unit tests) when present
task test                # 310 tests at 100% coverage
task frontend:test       # vitest incl. the new buildDagHash unit
task frontend:e2e        # Playwright against a real build

Checklist

• task lint passes (ruff format --check + ruff check).
• task test passes at 100% coverage (310 passed, TOTAL 100%).
• I followed the load-bearing patterns in .claude/design_patterns.md (extended an existing seam, didn't fork one). Added the "always-built artifact-derived data-dict section (Health Check)" pattern + its TOC entry in this PR.
• Data-dict / SPA changes keep producer ↔ consumer in sync (builder health key ⇄ data.js/service.js/ui.js) and don't re-inline the payload.
• Graph-UI changes rebuilt the committed bundle (task frontend:build) and I committed dbdocs/site/bundle/assets/graph/.
• New files under dbdocs/site/bundle/** (assets/css/icons.css) are covered by the pyproject.toml artifacts glob (dbdocs/site/bundle/**/*).
• Docs / dbdocs.yml.example updated for the new run_results/health knobs and the --run-results flag.

Screenshots / notes

• The Health Check works from an ordinary dbt build/dbt test — no extra dbt
  package. Without run_results.json, the manifest-derived dimensions still
  render; only the per-test pass/fail detail is skipped (a note records it).
• Status colors are consolidated onto --status-* CSS custom properties (one
  source of truth across pills/badges/scorecard, light + dark).
• Suggest adding before/after screenshots of the #/health page and a model
  page's Tests section here.